soil property estimates for the upper tana, kenya, derived ... · the information generated for the...

Report 2010/07

Soil property estimates for the Upper Tana, Kenya, derived from SOTER and WISE

(Version 1.0)

Niels H. Batjes

(May 2010)

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior permission of the copyright owner. Applications for such permission, with a statement of the purpose and extent of the reproduction, should be addressed to the Director, ISRIC - World Soil Information, PO B0X 353, 6700 AJ Wageningen, the Netherlands. The designations employed and the presentation of materials in electronic forms do not imply the expression of any opinion whatsoever on the part of ISRIC concerning the legal status of any country, territory, city or area or of is author concerning the delimitation of its frontiers or boundaries. Copyright © 2010, ISRIC - World Soil Information Disclaimer:

While every effort has been made to ensure that the data are accurate and reliable, ISRIC cannot assume liability for damages caused by inaccuracies in the data or as a result of the failure of the data to function on a particular system. ISRIC provides no warranty, expressed or implied, nor does an authorized distribution of the data set constitute such a warranty. ISRIC reserves the right to modify any information in this document and related data sets without notice. Correct citation: Batjes NH 2010. Soil property estimates for the Upper Tana, Kenya, derived from

SOTER and WISE (ver. 1.0). Report 2010/07, ISRIC – World Soil Information, Wageningen (41p. with data set)

(http://www.isric.org/isric/Webdocs/Docs/ISRIC_Report_2010_07.pdf)

Inquiries: c/o Director, ISRIC – World Soil Information PO Box 353 6700 AJ Wageningen The Netherlands E-mail: [email protected] Web: www.isric.org Front cover: Main FAO soil groupings for the Upper Tana, Kenya

http://www.isric.org/isric/Webdocs/Docs/ISRIC_Report_2010_07.pdf

mailto:[email protected]

http://www.isric.org/

i

Contents

Foreword........................................................................................iii

SUMMARY...................................................................................... iv

1 INTRODUCTION ........................................................................... 1

2 MATERIALS AND METHODS ............................................................ 2 2.1 Source of primary SOTER data .................................................. 2 2.2 Preparation of secondary SOTER data ........................................ 3

2.2.1 Checking of primary data................................................ 3 2.2.2 Filling gaps in the measured soil data ............................... 3 2.2.3 List of soil variables ....................................................... 6

3 RESULTS AND DISCUSSION ........................................................... 6 3.1 Map unit composition............................................................... 6 3.2 Soil property estimates ............................................................ 7 3.3 Type and number or taxotransfer rules used ............................... 9 3.4 Assumptions and limitations ................................................... 10 3.5 Linkage to GIS...................................................................... 11

4 CONCLUSIONS ........................................................................... 13

ACKNOWLEDGEMENTS ................................................................... 14

REFERENCES ................................................................................ 15

APPENDICES................................................................................. 19 Appendix 1. Structure of table SOTERunitComposition ..................... 19 Appendix 2. Structure of table SOTERparameterEstimates ................ 21 Appendix 3. Structure of table SOTERflagRules ............................... 22 Appendix 4. Structure of table SOTERsummaryFile .......................... 24 Appendix 5. Structure of table SOTERsummaryFile_Prop .................. 26 Appendix 6. Soil textural classes................................................... 27 Appendix 7. Installation............................................................... 28

ii

List of Tables

Table 1. FAO soil units mapped in SOTER_UT and number of similar soil profiles in

WISE used for taxotransfer rule development............................................. 4 Table 2. List of soil variables considered in secondary SOTER data sets ............... 5 Table 3. Criteria for defining confidence in the derived data ............................... 8 Table 4. Type and frequency of taxotransfer rules and expert rules applied .......... 9 Table 5. Conventions used for coding soil attributes in the taxotransfer scheme...22

List of Figures

Figure 1. Representation of SOTER units and their conceptual structure............... 2 Figure 2. Schematic representation of taxotransfer procedure for filling gaps

in SOTER .......................................................................................... 8 Figure 3. Schematized procedure for linking soil property estimates for the upper

layer of the main soil unit of a SOTER map unit with the geographical data ..12 Figure 4. Derived soil properties for dominant soil units in Upper Tana ...............13 Figure 5. SOTER soil texture classes ..............................................................27

Soil property estimates for the Upper Tana, Kenya iii

Foreword ISRIC – World Soil Information has the mandate to create and increase the awareness and understanding of the role of soils in major global issues. As an international institution, we inform a wide audience about the multiple roles of soils in our daily lives; this requires scientific analysis of sound soil information. This study presents derived soil properties for the Upper Tana, Kenya, for application in exploratory studies. It draws on two databases developed at ISRIC. First, the Soil and Terrain (SOTER) database for the Upper Tana, Kenya, at scale 1:250 000, compiled in the framework of the Green Water Credits Project. Being dependent on historic data, there are often gaps in the measured analytical data held in SOTER. ISRIC – World Soil Information has therefore developed a uniform, consistent methodology for filling common gaps in primary SOTER databases to produce secondary (SOTWIS) data sets for general-purpose applications. This procedure to derive secondary data, known as taxotransfer rule-based procedures, draws heavily on soil analytical data held in the ISRIC-WISE soil profile database. The consistent taxotransfer procedure has already been applied to SOTER datasets for Latin America and the Caribbean, Central and Eastern Europe, Southern Africa, Central Africa and other areas with SOTER-like databases. These secondary databases have been used in support of the ‘Harmonized World Soil Database’, a collaborative effort of Food and Agriculture Organization of the United Nations (FAO), International Institute for Applied Systems Analysis (IIASA), ISRIC - World Soil Information, Institute of Soil Science - Chinese Academy of Sciences (IISCAS), and Joint Research Centre of the European Commission (JRC). The information generated for the Upper Tana will be applied in a comprehensive hydrological model at the basin scale with land use practices for estimating the impact of soil-water conservation on soil erosion, agricultural productivity and water availability. In order to consolidate its world soil databases, ISRIC – World Soil Information is seeking collaboration with national institutes with a mandate for soil resource inventories. Dr Ir Prem Bindraban Director, ISRIC – World Soil Information

iv Soil property estimates for the Upper Tana, Kenya

SUMMARY

This report describes a harmonized set of soil property estimates for the Upper Tana, Kenya. The data set was derived from the 1:250 000 scale Soil and Terrain Database for the Upper Tana (SOTER_UT, ver. 1.0) and the ISRIC-WISE soil profile database, using standardized taxonomy-based pedotransfer (taxotransfer) procedures.

The land surface of the Upper Tana, Kenya, covering some 15,905 km2, has been mapped in SOTER using 186 unique SOTER units. Each map unit may comprise of up to four different soil components. In so far as possible, each soil component has been characterized by a regionally representative profile, selected and classified by national soil experts. Conversely, in the absence of any measured legacy data, soil components were characterized using synthetic profiles for which only the FAO-Unesco (1988) classification is known.

Soil components in SOTER_UT have been characterized using 144 profiles consisting of 108 real and 36 so-called synthetic profiles. The latter were used to represent some 18% per cent of the study area. Comprehensive sets of measured attribute data are seldom available for most profiles (108) collated in SOTER_UT, as these were not considered in the source materials. Consequently, to permit modelling, gaps in the soil analytical data have been filled using consistent taxotransfer procedures. Modal soil property estimates necessary to populate the taxotransfer procedure were derived from statistical analyses of soil profiles held in the ISRIC-WISE database. The current taxotransfer procedure only considers profiles in WISE that: (a) have FAO soil unit names (43) identical to those mapped for the Upper Tana in SOTER, and (b) originate from regions having similar Köppen climate zones (n= 5617).

Property estimates are presented for 18 soil variables by soil unit for fixed depth intervals of 0.2 m to 1 m depth: organic carbon, total nitrogen, pH(H2O), CECsoil, CECclay, base saturation, effective CEC, aluminium saturation, CaCO3 content, gypsum content, exchangeable sodium percentage (ESP), electrical conductivity (ECe), bulk density, content of sand, silt and clay, content of coarse fragments (> 2 mm), and volumetric water content (-33 kPa to -1.5 MPa). These attributes have been identified as being useful for agro-ecological zoning, land evaluation, crop growth simulation, modelling of soil carbon stocks and change, and studies of global environmental change.

Soil property estimates for the Upper Tana, Kenya v

The soil property estimates can be linked to the spatial data (map), using GIS, through the unique SOTER-unit code; database applications should consider the full map unit composition and depth range.

The derived data presented here may be used for exploratory assessments at basin scale (< 1:250 000). They should be seen as best estimates based on the current, still limited, selection of soil profiles in SOTER for the Upper Tana and data clustering procedure ― the type of taxotransfer rules used to fill gaps in the measured data has been flagged to provide an indication of confidence in the derived data.

Keywords: legacy soil data, taxotransfer procedures, derived soil properties, secondary data set, Upper Tana, Kenya, WISE database, SOTER database

vi Soil property estimates for the Upper Tana, Kenya

Soil property estimates for the Upper Tana, Kenya 1

ISRIC Report 2010/07

1 INTRODUCTION 1 ISRIC, FAO and UNEP, under the aegis of the International Union of Soil Sciences (IUSS), are updating the information on world soil resources in the World Soils and Terrain Digital Databases (SOTER) project. Once global coverage has been attained, a global SOTER is to supersede the 1:5 million scale Soil Map of the World (Nachtergaele and Oldeman 2002; Oldeman and van Engelen 1993).

SOTER databases are composed of two main elements: a geographic and an attribute data component. The geographical database holds information on the location, extent, and topology of each SOTER unit. The attribute database describes the characteristics of the spatial unit and includes both area data and point data. A geographical information system (GIS) is used to manage the geographic data, while the attribute data are handled in a relational database management system. Methodological details may be found in the SOTER Procedures Manual (van Engelen and Wen 1995).

Soil components of individual SOTER units are characterized by a representative soil profile (Figure 1). These legacy data are selected from available soil survey reports, as the SOTER program does not involve new ground surveys. As a result, there are often gaps in the measured (i.e. primary) analytical data, in particular the soil physical data. This precludes the direct use of primary SOTER data in models. ISRIC has therefore developed a uniform, consistent methodology for filling common gaps in primary SOTER databases to produce secondary (SOTWIS) data sets for general-purpose applications (Batjes 2003; Batjes et al. 2007). This taxotransfer rule-based procedure draws heavily on soil analytical held in the ISRIC-WISE soil profile database (Batjes 2009). So far, the consistent taxotransfer procedure has been applied to SOTER data for Latin America and the Caribbean, Central and Eastern Europe, Southern Africa, Central Africa and other areas with SOTER-like databases (see www.isric.org for details). The approach has also been used in support of the Harmonized World Soil Database (FAO/IIASA/ISRIC/ISSCAS/JRC 2009).

1 Note: Reports that describe secondary SOTER (SOTWIS) databases have similar structure and content, the main difference being the region-specific information presented in each document [NHB].

http://www.isric.org/UK/About+ISRIC/Projects/Track+Record/SOTER+data.htm

2 Soil property estimates for the Upper Tana, Kenya


Figure 1. Representation of SOTER units and their conceptual structure

This report discusses the application of the taxotransfer procedure to the primary SOTER data for The Upper Tana, Kenya (hereafter referred to as SOTER_UT). Chapter 2 describes the materials and methods with special focus on the procedure for preparing the secondary SOTER data. Results are discussed in Chapter 3, while concluding remarks are drawn in Chapter 4. The structure of the various output tables and installation procedure are documented in the Appendices.

2 MATERIALS AND METHODS

2.1 Source of primary SOTER data The SOTER database covering the Upper Tana, Kenya, compiled in the framework of the Green Water Credits Project (GWC), provided the basis for this study. The available soil geographical and attribute data were collated into SOTER format using the base materials described in Dijkshoorn et al. (2010). Although the map has a generalized scale of 1:250 000 million, the detail, and quality of the primary information varies widely within the study area.

All profiles in SOTER_UT were characterised according to the Revised Legend of FAO (1988) and World Reference Base for Soil Resources (FAO 2006). The Revised Legend, however, was used to display/map the soil units using GIS to ensure global consistency with earlier SOTER databases (e.g. FAO and ISRIC 2003; FAO et al. 1998).

http://www.isric.org/UK/About+ISRIC/Projects/Current+Projects/Green+Water+Credits.htm



2.2 Preparation of secondary SOTER data

2.2.1 Checking of primary data

The primary SOTER database was first screened for possible inconsistencies using automated integrity checks developed for WISE (Batjes 1995). All soil classifications were taken at face value; small inconsistencies in the analytical data, however, were corrected. The screened set provided the basis for the current analyses.

The screened dataset includes 144 so-called representative profiles, consisting of 108 real profiles, of which 100 are geo-referenced, and 36 virtual profiles. These profiles are physically linked to the spatial data in accord with SOTER standards.

In accord with SOTER conventions (van Engelen and Wen 1995), so-called virtual profiles have been introduced when the FAO classification for a given soil unit was known from soil maps for the region, but there are no real profiles (i.e. measured data) yet to characterize these units (see Dijkshoorn et al. 2010). For each virtual profile, the soil drainage class was inferred using expert judgement. [Note: Once parameter estimates for the virtual profiles have been estimated, using procedures described in this report, the profiles become so-called synthethic profiles]. Several map units consist of inland waters (KEns1) ― these are only characterized in the GIS-file.

2.2.2 Filling gaps in the measured soil data

Being based on available soil survey reports, there are always gaps in the soil analytical data ― the limited set of so-called “mandatory SOTER attributes” simply is not available for most profiles in SOTER_UT.

Gaps in the attribute data were filled here using consistent taxotransfer procedures (Batjes 2003; Batjes et al. 2007). The soil variables considered in the procedure are detailed in Section 2.3.3. The soil property estimates required to run these procedures were derived from statistical analyses of 5617 profiles extracted from version 3.1 of the ISRIC-WISE database (Batjes 2009). This selection only included those profiles in WISE that: (a) have similar FAO (1988) classification as mapped for SOTER_UT (Table 1),



and (b) originate from areas that have similar Köppen-Geiger climates as mapped for the Upper Tana by Kottek et al. (2006).

Table 1. FAO soil units mapped in SOTER_UT and number of similar soil profiles in WISE used for taxotransfer rule development

FAO soil units SOTER_UT a WISE b

Alh 1 (1/0) 60 Arc 2 (0/2) 55 Arh 3 (0/3) 240 Atc 1 (1/0) 5 CLh 1 (1/0) 152 CLl 5 (5/0) 44 CLp 7 (6/1) 54 CMc 6 (5/1) 176 Cme 2 (1/1) 205 CMg 1 (0/1) 78 Cmu 1 (1/0) 69 CMv 3 (3/0) 64 CMx 2 (1/1) 68 FLc 4 (2/2) 130 Fle 1 (0/1) 133 FLs 1 (0/1) 11 Gle 3 (0/3) 163 GLk 1 (0/1) 13 Gyh 3 (1/2) 11 Gyk 2 (0/2) 21 Gyp 3 (1/2) 11 KSh 1 (0/1) 25 KSk 3 (3/0) 52 Lpe 3 (0/3) 70 LPk 2 (0/2) 33 LVj 1 (1/0) 58 LVk 3 (3/0) 127 LVv 1 (1/0) 34 LXh 2 (2/0) 199 PHl 1 (1/0) 129 PLd 1 (1/0) 31 PZh 1 (1/0) 68 RGc 5 (3/2) 97 Rge 3 (1/2) 164 Rgy 3 (0/3) 5 SCg 2 (0/2) 35 SCn 2 (1/1) 33 Scy 3 (1/2) 33 SNh 1 (0/1) 96 SNk 1 (1/0) 48 Vre 2 (2/0) 311



SOTER_UT a WISE bFAO soil units

VRk 4 (4/0) 148 Vry 1 (0/1) 7

a First number is for total number of soil profiles linked to the SOTER-GIS map units; the first number in brackets is for measured profiles, the second for virtual profiles (i.e. profiles for which there are no measured data; these have codes like KEsyn15 or KEsynLPe)

b Number of profiles from WISE considered in the taxotransfer scheme (n = 5617); for details see text.

Measured values in WISE that underlie the taxotransfer scheme — like those held in SOTER_UT — will reflect both variations inherent to the soil unit and those that can be ascribed to the methods of sampling and measurement. For reasons outlined earlier (Batjes 2002, p. 6-11), a pragmatic approach to the comparability of soil analytical data had to be adopted for use with small scale SOTER databases. A similar approach has also been used with the Harmonized World Soil Database (FAO/IIASA/ISRIC/ISSCAS/JRC 2009). Although this type of approach is Table 2. List of soil variables considered in secondary SOTER data sets

Organic carbon Total nitrogen Soil reaction (pHH2O) Cation exchange capacity (CECsoil) Cation exchange capacity of clay size fraction (CECclay)

a b

Base saturation (as % of CECsoil) b

Effective cation exchange capacity (ECEC) b c Aluminium saturation (as % of ECEC) b CaCO3 content Gypsum content Exchangeable sodium percentage (ESP) b

Electrical conductivity (ECe) Bulk density Coarse fragments (> 2 mm, volume %) Sand (mass %) Silt (mass %) Clay (mass %) Available water capacity (cm3 cm-3 102 or vol%; -33 kPa to -1.5 MPa) b d

a CECclay was calculated from CECsoil by assuming a mean contribution of 350 cmolc per 100 g

OC, the common range being from 150 to over 750 cmolc per 100 g (Klamt and Sombroek 1988). Similarly, as a rule of thumb, CECOC values of 300 to 400 cmolc per 100 g OC (NH4OAc, pH 7.0), are used by USDA-NRCS (1995 p. 26).

b Calculated from other measured soil properties. c ECEC is defined here as exchangeable (Ca+++ Mg+++ K++ Na+) + exchangeable (Al+++) in

accordance with USDA-NRCS (1995); see also FAO (2006, p. 125). d Limits for soil water potential for Available Water Capacity (AWC) conform to USDA

standards (Soil Survey Staff 1983); these values are not corrected for volume percentage of coarse fragments.



considered appropriate for soil data applications at broad scale, correlation of soil analytical data should be done more rigorously when more precise scientific research is considered.

The analytical data for each combination of soil unit, texture class and depth layer were screened using a robust outlier scheme, by attribute (see Batjes 2003). The output of the taxotransfer procedure has been stored in a secondary data set (known as SOTWIS database); for details see Appendix 1.

2.2.3 List of soil variables

Special attention has been paid to those key attributes (Table 2) that are commonly required in studies of agro-ecological zoning, food productivity, soil gaseous emissions/sinks and environmental change (see Batjes et al. 1997; Bouwman et al. 2002; Cramer and Fischer 1997; Easter et al. 2007; FAO/IIASA/ISRIC/ISSCAS/JRC 2009; Fischer et al. 2002; Scholes et al. 1995).

Table 2 does not include soil hydraulic properties because measured data for the latter are generally lacking in the systematic soil survey reports that underlie SOTER and WISE.

3 RESULTS AND DISCUSSION

3.1 Map unit composition

The Upper Tana has been characterized using 190 unique map units or SOTER units in the SoilComponent table (Figure 1); these comprise 192 terrain components and 262 soil components. It contains data for four SOTER units that are not shown on the GIS map: KE237, KE239, KE240, and KE411. As a result, the GIS map considers data for 186 unique map units, corresponding with 617 polygons on the map. Some 18% of the study area has been characterized by a synthetic profile. At the small scale under consideration, many mapping units are compound; they may consist of up to four different soil units. This map unit complexity must be considered when using the data; typically, this will



have to be done using software, specifically written for a particular application (e.g. Batjes et al. 2007; Easter et al. 2007). Overall, the Upper Tana has been characterized using 43 different soil units (FAO Revised Legend).

The full composition of each SOTER unit has been summarized in table SOTERunitComposition (Appendix 1). This table lists the name and relative area of the main major FAO soil group for each map unit, as well as the type and relative area of all the component soil units.

3.2 Soil property estimates

The taxotransfer procedure generates soil property estimates for five standardized depth ranges of 20 cm each to 1 m, and 2 standardized depth ranges of 50 cm (100-150 cm and 150-200 cm) (see Batjes 2008). Inherently, property estimates for the deeper layers are considered less reliable than those for the upper layers of soil as they are based on less extensive data sets. Therefore, the current data set only presents derived data up to 1 m depth, or less when applicable (e.g. for shallow Leptosols).

In case of missing measured values in SOTER, the cut-off point for applying any taxotransfer rule is nWISE < 5; that is there should be at least 5 cases in the WISE subset for the corresponding combination of soil unit, soil variable, soil layer, and soil textural class in order to apply the substitution procedure. Soil textural classes were defined in accordance with current SOTER standards – coarse, medium, fine, very fine and medium fine (Figure 4, Appendix 6). The taxotransfer procedure is summarized in Figure 2; see also Appendix 3.

Each flag listed under TTRsub (where sub stands for FAO soil unit) and TTRmain (where main stands for major soil group) consists of a sequence of letters followed by a numeral, for example A3h2. The letters indicate soil attributes for which a taxotransfer rule has been applied; coding conventions are explained in Appendix 3. The number code reflects the size of the sample population in WISE, after outlier rejection, on which the statistical analyses that underlie taxotransfer scheme were based (Table 3).



CLAF PRID LAYER Newtopdep Newbotdep TTRsub TTRmain CMd KEhyp04 D1 0 18 b3c2j3o3r2 a2h1

CMd KEhyp04 D1 18 20 C3j1 A3h2

Soil property estimates based on WISE-derived data, using data for the corresponding major soil group and either the same textural class (small letter) or undifferentiated textural class (capital).

Soil property estimates based on WISE-derived data, using data for the corresponding soil unit and same textural class: - b: Base saturation, 3 ( nWISE = 5 –14) - c: Bulk density, 2 ( nWISE = 15 – 29) - j: Exchangeable sodium percentage, 3 (nWISE = 5 –14) - o: Volumetric water content, 3 ( nWISE = 5 –14) - r: Total Nitrogen, 2 (nWISE = 15 – 29)

Figure 2. Schematic representation of taxotransfer procedure for filling gaps in SOTER

When a small letter is used for TTRsub, the substitution was based on median data for the corresponding soil unit, depth layer and textural class (for example, Rhodic Ferralsols (FRr), 0-20 cm (D1), Fine and nWISE > 5). Otherwise, when a capital is used, this indicates that the substitution for the given soil attribute was based on the whole set for the corresponding soil unit and depth layer, irrespective of soil texture (i.e. undifferentiated or u). The same coding conventions apply for TTRmain, but substitutions then consider derived soil data for the corresponding major FAO soil group.

Table 3. Criteria for defining confidence in the derived data Code Confidence level nWISE

a

1 Very high > 30 2 High 15-29 3 Moderateb 5-14 4 Low 1-4 - No data 0 a nWISE is the sample size after the screening or outlier rejection procedure b The cut-off point in the TTR-approach is nWISE < 5



Expert rules are applied after the taxotransfer rules to remedy possible pedological inconsistencies (or artefacts) that may have arisen in the TTR-derived data. Such a check is necessary because individual TTR-rules do not consider possible correlations between different soil variables. For example, one expert rule (XR-TCEQ) checks whether there are indeed no carbonates in acid soil layers. Similarly, another expert rule (XR-BSAT) checks whether base saturation is low in acid soils and so on. In view of the diversity of soils worldwide, however, it remains difficult to account for all possible situations; this should be kept in mind when combining various derived data.

Derived soil data, resulting from the taxotransfer procedure, are presented in table SOTERparameterEstimates; see Appendix 2 for details.

3.3 Type and number or taxotransfer rules used There are numerous gaps in the primary soil analytical data in SOTER_UT (see 2.2.1). Table 4 lists how often each taxotransfer respective or expert-rule has been applied for each attribute as a percentage of the total number of “horizon/layer/depth” combinations in the secondary SOTER or SOTWIS set; details may be found in table SOTERflagTTRrules (Appendix 3). Table 4 shows, for example, that available water capacity (AWC) has been estimated in 89% of the cases, either using data for similar soil units (74% of cases, see under TTRsub) resp. similar major soil groups (15% of cases, see under TTRmain). Further, expert rules for available (XR-AWC) have been applied in 3% of the cases. This shows that in 92% of the cases, AWC for a given profile and fixed-depth layer, had to be estimated in this study due to the limited availability of measured water retention data for the Upper Tana region in SOTER. Table 4. Type and frequency of taxotransfer rules (TTR) and expert rules (XR) applied

Frequency of occurrence (%) TTR code (SOTNAM) TTRsub TTRmain TTR total Expert rules

TTR-BSAT 28 0 28 - TTR-BULK 64 3 67 - TTR-CECC 37 0 37 - TTR-CECS 16 0 16 - TTR-GRAV 0 0 0 - TTR-CLAY 16 0 16 - TTR-ECEC 79 0 79 -



Frequency of occurrence (%) TTR code (SOTNAM) TTRsub TTRmain TTR total Expert rules

TTR-ELCO 14 6 20 - TTR-ESP 26 0 26 - TTR-GYPS 7 11 18 - TTR-PHAQ 15 0 15 - TTR-SAND 16 0 16 - TTR-SILT 16 0 16 - TTR-TAWC 74 15 89 - TTR-TCEQ 19 4 23 - TTR-TOTC 34 0 34 - TTR-TOTN 92 1 93 - XR0-Text - - - 16 XR1-Alsa - - - 73 XR2-Bsat - - - 4 XR3-Elco - - - 48 XR4-Gyps - - - 21 XR5-CaCo - - - 27 XR6-CECc - - - 5 XR7-ESP - - - 0 XR8-CFRA - - - 0 XR9-BULK - - - 0 XR10-AWC - - - 3 Note: For definitions of abbreviations see text and Table 4, see also Appendix 3; ‘-‘ stands for not applicable.

3.4 Assumptions and limitations

Soil unit classifications (FAO 1988), as presented in the primary SOTER_UT database, were taken at face value. Soil experts, however, may classify the same soil profile differently when the available soil morphological and soil analytical data are ‘limited’ and subjective assumptions have to be made (e.g., Goyens et al. 2007; Kauffman 1987; Spaargaren and Batjes 1995). The soil classification code, however, is the primary driver of the taxotransfer procedure (see 2.2.2).

The overall assumption has been that the confidence in a TTR-based property estimate should increase with the size of the corresponding sample populations present in WISE, for the relevant soil units and Köppen climate zones, after outlier-rejection. In addition, the confidence in soil property estimates listed under TTRsub should be higher than for those listed under TTRmain resp. “derived using expert rules.”



A high confidence rating for a given property estimate, however, does not necessarily imply that this estimate will be representative for the soil unit under consideration. Profile selection for SOTER and WISE, as for many other small scale soil databases, is not probabilistic but based on available data and expert knowledge. Several of the soil attributes under consideration in Table 2 are not diagnostic in the Revised Legend (FAO 1988). In addition, some soil properties are readily modified by changes in land use or management, for example soil pH, aluminium saturation, soil salinity, and organic matter content. Information on land use/management history by profile, however, is seldom available in SOTER and, as such, this aspect could not be considered explicitly in the taxotransfer procedure yet.

Finally, it should be noted that adoption of different criteria for clustering data would inherently lead to varying property estimates. For example, selecting a different soil classification system (e.g., FAO 1974, FAO 1988 or WRB 2006), limits for depth layers (e.g., 0-20 cm intervals up to 100 cm versus 0-30 cm and 30-100 cm), criteria for defining soil textural classes (e.g., 5 classes in SOTER versus 3 classes for the FAO Soil Map of the World), choice of critical limits for applying taxotransfer rules (i.e. reject when nWISE < 5 or nWISE < 15), as well as the type of outlier-rejection and statistical procedures used, and the number of WISE profiles under consideration. Most importantly, however, the outcome will be determined primarily by the number and quality of the profile data collated in the underpinning, primary SOTER database. In particular, their geographic distribution over the region respectively various SOTER units, the degree to which the various data-fields have been filled, and the overall comparability of analytical methods used.

3.5 Linkage to GIS

SOTER units mapped for the region comprise up to four soil components. The full map unit composition has been summarized in one single table (SOTERunitComposition, see Appendix 1). Results of the taxotransfer procedure for each soil component, as typified by the representative profile, are stored in table SOTERparameterEstimates (Appendix 2). Results in this table have been linked to the corresponding SOTER units in two tables having the same content, but different data structures: a) SOTERsummaryFile, in which data by layer (Di) are presented vertically by NEWSUID, TCID and SCID (Appendix 4), and b) SOTERsummaryFile_Prop in which derived data for layer D1 to D5 are data presented horizontally by NEWSUID, TCID and SCID (Appendix 5).



Figure 3. Schematized procedure for linking soil property estimates for the upper layer (D1) of the main soil unit (TCID=1; SCID=1) of a SOTER map unit with the geographical data

Data in the later tables can be linked to GIS through NEWSUID, the unique SOTER map unit code. The overall procedure is visualized in Figure 3 for a hypothetical database. It should be noted here that GIS can only be used to display one “set of attributes” at a time per polygon or SOTER map unit. As an example, derived topsoil properties for organic carbon content, bulk density and available water capacity for the dominant soil component in a SOTER unit (i.e. TCID=1, SCID=1 and Layer= D1) are shown in Figure 4; classification is according to natural breaks (Jenks).

Typically, specific data selections that consider the full soil unit composition of individual SOTER units will have to be made before “aggregated” model output can be coupled back to the mapping units in the GIS. Details of such an approach may be found in Easter et al. (2007).



Figure 4. Derived soil properties for dominant soil units in Upper Tana

4 CONCLUSIONS

• The detail and quality of primary soil and terrain data underpinning SOTER_UT resulted in a variable resolution of the secondary product presented here.

• Linkage between the soil profile data and the spatial component of SOTER_UT required generalisation of measured soil (profile) data by soil unit and depth zone. This involved the transformation of variables that show a marked spatial and temporal variation and that have been determined in a range of laboratories, according to various analytical methods.

• A pragmatic approach to the comparability of soil analytical data has been adopted when developing the taxotransfer procedure. Although this is considered acceptable at the present broad scale, such a comparison must be done more rigorously when more detailed scientific work is considered.



• The derived soil data presented here can be used for exploratory assessments at subnational scale ― they should be seen as best estimates based on the current selection of soil profiles in SOTER_UT and data clustering procedure. Once additional profiles become available for the region in SOTER format, the present set of derived soil data should be refined.

• End-users should familiarize themselves with the procedures and assumptions that have been used to derive the soil property estimates prior to using them in models ― possible uncertainties are documented in the data set.

ACKNOWLEDGEMENTS Special thanks are due to my colleague Koos Dijkshoorn for useful discussions concerning the primary SOTER database for the Upper Tana and for liaising with staff at the Kenya Soil Survey (KSS), Kenya Agricultural Research Institute (KARI), concerning the primary SOTER data in the context of the Green Water Credits Project. Green Water Credits is supported by the International Fund for Agricultural Development. The proof of concept was part-financed by the Swiss Agency for Development and Cooperation. The program is implemented by an international consortium that includes: ISRIC – World Soil Information, Stockholm Environmental Institute, International Institute for Environment and Development, and Agricultural Economics Research Institute.

http://www.isric.org/isric/webdocs/docs/GWC_PolicyBrief12web.pdf



REFERENCES Batjes NH 1995. World Inventory of Soil Emission Potentials: WISE 2.1 -

Profile database user manual and coding protocols. Tech. Pap. 26, ISRIC, Wageningen

Batjes NH 2002. Soil parameter estimates for the soil types of the world for use in global and regional modelling (Version 2.1). ISRIC Report 2002/02c, International Food Policy Research Institute (IFPRI) and International Soil Reference and Information Centre (ISRIC), Wageningen. Available at: http://www.isric.org/isric/webdocs/Docs/ISRIC_Report_2002_02c.pdf; accessed 12/2008.

Batjes NH 2003. A taxotransfer rule-based approach for filling gaps in measured soil data in primary SOTER databases (GEFSOC Project). Report 2003/03, ISRIC - World Soil Information, Wageningen (http://www.isric.org/isric/webdocs/Docs/ISRIC_Report_2003_03.pdf)

Batjes NH 2008. Mapping soil carbon stocks of Central Africa using SOTER. Geoderma 146, 58-65

Batjes NH 2009. Harmonized soil profile data for applications at global and continental scales: updates to the WISE database. Soil Use and Management 25, 124-127 (http://dx.doi.org/10.1111/j.1475-2743.2009.00202.x )

Batjes NH, Fischer G, Nachtergaele FO, Stolbovoy VS and van Velthuizen HT 1997. Soil data derived from WISE for use in global and regional AEZ studies (ver. 1.0). Interim Report IR-97-025, FAO/ IIASA/ ISRIC, Laxenburg (http://www.iiasa.ac.at/Admin/PUB/Documents/IR-97-025.pdf)

Batjes NH, Al-Adamat R, Bhattacharyya T, Bernoux M, Cerri CEP, Gicheru P, Kamoni P, Milne E, Pal DK and Rawajfih Z 2007. Preparation of consistent soil data sets for SOC modelling purposes: secondary SOTER data sets for four case study areas. Agriculture, Ecosystems and Environment 122, 26-34

Bouwman AF, Boumans LJM and Batjes NH 2002. Modeling global annual N2O and NO emissions from fertilized fields. Global Biogeochemical Cycles 16, 1080, doi:10.1029/2001GB001812

CEC 1985. Soil Map of the European Communities (1:1,000,000). Report EUR 8982, Office for Official Publications of the European Communities, Luxembourg

Cramer W and Fischer A 1997. Data requirements for global terrestrial ecosystem modelling. In: Walker and Steffen (editors), Global Change and Terrestrial Ecosystems. Cambridge University Press, Cambridge, pp 529-565

Dijkshoorn JA, Machiaria P, Huting JRM and Maingi p 2010. Soil and and terrain conditions for the Upper Tana, Kenya, Kenya Soil Survey (KSS)

http://www.isric.org/isric/webdocs/Docs/ISRIC_Report_2002_02c.pdf;

http://www.isric.org/isric/webdocs/Docs/ISRIC_Report_2003_03.pdf

http://dx.doi.org/10.1111/j.1475-2743.2009.00202.x

http://dx.doi.org/10.1111/j.1475-2743.2009.00202.x

http://www.iiasa.ac.at/Admin/PUB/Documents/IR-97-025.pdf

http://www.iiasa.ac.at/Admin/PUB/Documents/IR-97-025.pdf



/Kenya Agricultural Research Institute (KARI) and ISRIC – World Soil Information, Wageningen (in prep., version May 2010)

Easter M, Paustian K, Killian K, Williams S, Feng T, Al-Adamat R, Batjes NH, Bernoux M, Bhattacharyya T, Cerri CC, Cerri CEP, Coleman K, Falloon P, Feller C, Gicheru P, Kamoni P, Milne E, Pal DK, Powlson D, Rawajfih Z, Sessay M and Wokabi S 2007. The GEFSOC soil carbon modeling system: a tool for conducting regional-scale soil carbon inventories and assessing the impacts of land use change on soil carbon. Agriculture, Ecosystems & Environment 122, 13-25

FAO 1988. FAO-Unesco Soil Map of the World, Revised Legend, with corrections and updates. World Soil Resources Report 60, FAO, Rome; reprinted with updates as Technical Paper 20 by ISRIC, Wageningen,1997

FAO 2006. World Reference Base for soil resources - A framework for international classification, correlation and communication. World Soil Resources Reports 103, International Union of Soil Sciences, ISRIC - World Soil Information and Food and Agriculture Organization of the United Nations, Rome (http://www.fao.org/ag/agl/agll/wrb/doc/wrb2007_corr.pdf)

FAO and ISRIC 2003. Soil and Terrain database for Southern Africa (1:2 million scale). FAO Land and Water Digital Media Series 25, ISRIC and FAO, Rome

FAO, ISRIC, UNEP and CIP 1998. Soil and terrain digital database for Latin America and the Caribbean at 1:5 million scale. Land and Water Digital Media Series No. 5, Food and Agriculture Organization of the United Nations, Rome

FAO/IIASA/ISRIC/ISSCAS/JRC 2009. Harmonized World Soil Database (version 1.1), Prepared by Nachtergaele FO, van Velthuizen H, Verelst L, Batjes NH, Dijkshoorn JA, van Engelen VWP, Fischer G, Jones A, Montanarella L., Petri M, Prieler S, Teixeira E, Wiberg D and Xuezheng Shi. Food and Agriculture Organization of the United Nations (FAO), International Institute for Applied Systems Analysis (IIASA), ISRIC - World Soil Information, Institute of Soil Science - Chinese Academy of Sciences (ISSCAS), Joint Research Centre of the European Commission (JRC), Laxenburg, Austria (available at http://www.iiasa.ac.at/Research/LUC/luc07/External-World-soil-database/HTML/index.html?sb=1 )

Fischer G, van Velthuizen HT, Shah M and Nachtergaele FO 2002. Global Agro-ecological Assessment for Agriculture in the 21st Century: Methodology and Results. RR-02-02, International Institute for Applied Systems Analysis (IIASA) and Food and Agriculture Organization of the United Nations (FAO), Laxenburg

Goyens C, Verdoodt A, Van de Wauw J, Baert G, van Engelen VWP, Dijkshoorn JA and Van Ranst E 2007. Base de données numériques sur les sols et le terrain (SOTER) de l'Afrique Centrale (RD Congo, Rwanda et Burundi). Etude et Gestion des Sols 14, 207-218

http://www.fao.org/ag/agl/agll/wrb/doc/wrb2007_corr.pdf

http://www.iiasa.ac.at/Research/LUC/luc07/External-World-soil-database/HTML/index.html?sb=1

http://www.iiasa.ac.at/Research/LUC/luc07/External-World-soil-database/HTML/index.html?sb=1



Kauffman JH 1987. Comparative classification of some deep, well-drained red clay soils of Mozambique. Technical Paper 16, International Soil Reference and Information Centre (ISRIC), Wageningen. Aavailable at: http://www.isric.org/isric/webdocs/Docs/ISRIC_TechPap16.pdf; accessed June 2008.

Klamt E and Sombroek WG 1988. Contribution of organic matter to exchange properties of Oxisols. In: Beinroth, Camargo and Eswaran (editors), Classification, characterization and utilization of Oxisols. Proc. of the 8th International Soil Classification Workshop (Brazil, 12 to 23 May 1986). Empresa Brasileira de Pesquisa Agropecuaria (EMBRAPA), Soil Management Support Services (SMSS) and University of Puerto Rico (UPR), Rio de Janeiro, pp 64-70

Kottek M, Grieser J, Beck C, Rudolf B and Rubel F 2006. World Map of the Köppen-Geiger climate classification updated. Meteorol. Z. 15, 259-263 ( DOI: 10.1127/0941-2948/2006/0130)

Landon JR 1991. Booker Tropical Soil Manual. Longman Scientific & Technical, New York, 474 p

Marsman BA and de Gruijter JJ 1986. Quality of soil maps: A comparison of soil survey methods in sandy areas. Soil Survey Papers 15, Netherlands Soil Survey Institute, Wageningen

Nachtergaele FO and Oldeman LR 2002. World soil and terrain database (SOTER): past, present and future. Transactions 17th World Congress of Soil Science. International Union of Soil Sciences (IUSS), Bangkok, pp 653/1-653/10

Oldeman LR and van Engelen VWP 1993. A World Soils and Terrain Digital Database (SOTER) - An improved assessment of land resources. Geoderma 60, 309-335

Scholes RJ, Skole D and Ingram JS 1995. A global database of soil properties: proposal for implementation. IGBP-DIS Working Paper 10, International Geosphere Biosphere Program, Data & Information System, Paris

Soil Survey Staff 1983. Soil Survey Manual (rev. ed.). United States Agriculture Handbook 18, USDA, Washington (Available through: http://soils.usda.gov/technical/manual/; Accessed: June 2008)

Spaargaren OC and Batjes NH 1995. Report on the classification into FAO-Unesco soil units of profiles selected from the NRCS pedon database for IGBP-DIS. Work. Pap. 95/01, ISRIC, Wageningen

USDA-NRCS 1995. Soil Survey Laboratory Information Manual. Soil Survey Investigations Report No. 45 (version 1.0, May 1995) [Available at ftp://ftp-fc.sc.egov.usda.gov/NSSC/Lab_Info_Manual/ssir45.pdf ; verified 11 January 2007], National Soil Survey Center, Soil Survey Laboratory Lincoln (Nebraska)

http://www.isric.org/isric/webdocs/Docs/ISRIC_TechPap16.pdf;

http://soils.usda.gov/technical/manual/;

ftp://ftp-fc.sc.egov.usda.gov/NSSC/Lab_Info_Manual/ssir45.pdf



van Engelen VWP and Wen TT 1995. Global and National Soils and Terrain Digital Databases (SOTER): Procedures Manual (rev. ed.). (Published also as FAO World Soil Resources Report No. 74), UNEP, IUSS, ISRIC and FAO, Wageningen



APPENDICES

Appendix 1. Structure of table SOTERunitComposition Table SOTERunitComposition, in MS-Access® format, gives the full composition of each SOTER unit in terms of its: landform, lithology (parent material), dominant major FAO soil group and its relative extent, then component in soil units with their relative extent, and the identifier for the corresponding representative profile. The relevant information was distilled from three primary SOTER tables, viz. Terrain, SoilComponent, and Profile, to facilitate data processing. The content of this table can be linked to the geographical data in a GIS through the unique SOTER unit code or NEWSUID, a combination of the fields for ISO and SUID. Structure of table SOTERunitCompositiona

Name Type Description

ISOC Text ISO-3166 country code (1994) or WD for World SUID Integer The identification code of a SOTER unit on the map and in the database NEWSUID Text Globally unique code for SOTER unit, comprising fields ISOC plus SUID (e.g. KE0115) LNDF Text Code for SOTER landforms (see SOTWIS_codes) LITH Text Code for SOTER lithology (See SOTWIS_codes) NoOfSoilComp Text Number of soil components in given SOTER unit DomFAOgroup Text Dominant FAO major soil group in SOTER (Note: This

need not always be SOIL1) PropDomFAOGroup Proportion of dominant major soil group in SOTER unit (%) PropSynthProf Proportion of SOTER unit characterized by a synthethic

profile (%) SoilMapunit Text Aggregated code for map unit summarizing the overall composition b SOIL1 Text Characterization of the first (main) soil unit according to the Revised FAO-Unesco Legend PROP1 Integer Proportion, as a percentage, that the main soil unit occupies within the SOTER unit PRID1 Text Unique code for the corresponding measured resp. virtual soil profile (e.g. KN_LPe_syn) SOIL2 Text As above but for the next soil unit PROP2 Integer As above PRID2 Text As above SOIL3 Text As above but for the next soil unit PROP3 Integer As above PRID3 Text As above




SOIL4 Text As above but for the next soil unit PROP4 Integer As above PRID4 Text As above SOIL5 Text As above but for the next soil unit PROP5 Integer As above PRID5 Text As above SOIL6 Text As above but for the next soil unit PROP6 Integer As above PRID6 Text As above SOIL7 Text As above but for the next soil unit PROP7 Integer As above PRID7 Text As above SOIL8 Text As above but for the next soil unit PROP8 Integer As above PRID8 Text As above SOIL9 Text As above but for the next soil component PROP9 Integer As above PRID9 Text As above SOIL10 Text As above but for the next soil component PROP10 Integer As above PRID10 Text As above

a Generally, not all 10 available fields for SOILi will be filled in SOTER. In the case of The Upper Tana, Kenya, from 1 up to 4 different soil components have been defined for each map unit. b These codes have the following format: VRe2GLe4. The relative extent of each soil unit (e.g., VRe) has been expressed in 5 classes to arrive at a compact map unit code: 1 – from 80 to 100 per cent; 2 – from 60 to 80 per cent; 3 – from 40 to 60 percent; 4 – from 20 to 40 per cent, and 5 – less than 20 percent.



Appendix 2. Structure of table SOTERparameterEstimates Table SOTERparameterEstimates lists property estimates ― depth-weighted by layer ― for all soil units (represented by their PRID) that have been mapped for the study region. This information can be linked to the soil geographical data – in a GIS – through the unique profile code (PRID). Structure of table SOTERparameterEstimates


CLAF Text FAO-Unesco (1988) Revised Legend code PRID Text profile ID (as listed in SOTERmapunitComposition) Drain Text FAO soil drainage class Layer Text code for depth layer (from D1 to D5; e.g. D1 is from 0 to 20 cm etc.) TopDep Integer depth of top of layer (cm) BotDep Integer depth of bottom of layer (cm) CFRAG Integer coarse fragments (vol.% > 2 mm) SDTO Integer sand (mass %) STPC Integer silt (mass %) CLPC Integer clay (mass %) PSCL Text SOTER texture class (see Appendix 6) BULK Single bulk density (kg dm-3) TAWC Integer available water capacity (cm3 cm-3 102, -33 kPa to -1.5 MPa conform to USDA standards) CECs Single cation exchange capacity (cmolc kg-1) for fine earth fraction BSAT Integer base saturation as percentage of CECsoil

ESP Integer exchangeable Na as percentage of CECsoil CECc Single CECclay, corrected for contribution of organic matter (cmolc kg-1) a PHAQ Single pH measured in water TCEQ Single total carbonate equivalent (g C kg-1) GYPS Single gypsum content (g kg-1) ELCO Single electrical conductivity (dS m-1) TOTC b Single organic carbon content (g C kg-1) TOTN Single total nitrogen (g N kg-1) ECEC Single effective CEC (cmolc kg-1) ALSA Integer exchangeable Aluminium as percentage of ECEC

a CECclay is only calculated for layers where clay content >5%; else CECclay is set at -9 (see

Appendix 3). b Please note that TOTC is a field name used in SOTER representing organic carbon content

only, not total carbon!

Contents of table SOTERparameterEstimates should be consulted in conjunction with table SOTERflagTTRrules. The later lists the taxotransfer rules that have been applied for each profile, by depth layer and soil attribute. Details are given in Appendix 3.



Appendix 3. Structure of table SOTERflagRules Table SOTERflagTTRrules documents the type of taxotransfer rules that have been used to create table SOTERparameterEstimates (Appendix 2). Coding conventions are detailed in Table 5. Structure of table SOTERflagTTRrules


CLAF Text FAO Legend code PRID Text Unique identifier for representative profile Layer Text code for depth layer (from D1 to D5; e.g. D1 is from 0 to 20 cm) Newtopdep Integer Depth of top of layer (cm) Newbotdep Integer Depth of bottom of layer (cm) TTRsub Text Code showing the type of taxotransfer rule used (based on derived data for soil units; see text) TTRmain Text Code showing the type of taxotransfer rule used (based on derived data for major units; see text) TTRexpert Text Additional flags (based on expert-rules)

Note: Expert rules (TTRexpert) are run after the TTR-procedures (see text). For example, exchangeable aluminium percentage (ALSA) has been set at zero when pHwater is higher than 5.5. Similarly, the content of gypsum (GYPS) and content of carbonates (TCEQ) have been set at zero when pHwater is less than 6.5. Finally, the CEC of the clay fraction (CECclay) has been re-calculated from the depth-weighted measured and TTR-derived data for CECsoil and content of organic carbon assuming a mean contribution of 350 cmolc kg-1 OC, the common range being from 150 to over 750 cmolc per 100 g (Klamt and Sombroek 1988) ― CECclay values presented here thus are only rough estimates. Table 5. Conventions used for coding soil attributes in the taxotransfer scheme TTRflag SOTnam WISnam SoilVariable Comments

A ALSA ALSA ALSAT Exch. Aluminium percentage (% of ECEC)

B BSAT BSAT BSAT base saturation (% of CECs)

C BULK BULK BULKDENS Bulk density

D CECC CECC CECCLAY cation exchange capacity of clay fraction

E CECS CECS CECSOIL cation exchange capacity

F CFRAG GRAV GRAVEL coarse fragments

G CLPC CLAY CLAY clay %

H ECEC ECEC ECEC Effective CEC

I ELCO ECE ECE electrical conductivity

J ESP ESP ESP exchangeable Na percentage (% of CECs)



TTRflag SOTnam WISnam SoilVariable Comments

K GYPS GYPS GYPSUM gypsum content (g kg-1)

L PHAQ PHH2 PHH2O pH in water

M SDTO SAND SAND sand %

N STPC SILT SILT silt %

O AWC AWC AWC Vol. water content (-33 kPa to -1.5 MPa)

P TCEQ CACO CACO3 carbonate content (g kg-1)

Q TOTC ORGC ORGC organic carbon content (g C kg-1)

R TOTN TOTN TOTN total nitrogen content (g N kg-1)

Y --- --- --- PSCL estimated from TTR-derived sand, silt and clay content (where applicable)

Abbreviations: TTRflag = code for TTR-rule; SOTnam = codes used in SOTER; WISnam= codes used in WISE; SoilVariable= soil variables as described in Table 2 (page 8).



Appendix 4. Structure of table SOTERsummaryFile Table SOTERsummaryFile has been created to facilitate access to the derived data. For each SOTER unit (NEWSUID) on the map, it lists the soil property estimates by component soil unit and depth layer. Layer data are presented in one single column, i.e. vertically (see also Appendix 5). Structure of table SOTERsummaryFile


ISOC Text ISO-3166 country code (1994) SUID Integer The identification code of a SOTER on the map and in the database NEWSUID Text Globally unique map unit code, comprising fields ISOC plus SUID TCID Integer Number of terrain component in given map unit SCID Integer Number of soil unit within the given SOTER unit Layer Text Code for depth layer (from D1 to D7; e.g., D1 is from 0 to 20 cm and D7 from 150 to 200 cm) PROP Integer Relative proportion of SCID in given SOTER unit CLAF Text FAO-Unesco Revised Legend code PRID Text Profile ID (see table SOTERunitComposition) Drain Text FAO soil drainage class TopDep Integer Upper depth of layer (cm) BotDep Integer Lower dept of layer (cm) CFRAG Integer Coarse fragments (vol. % > 2 mm) SDTO Integer Sand (mass %) STPC Integer Silt (mass %) CLPC Integer Clay (mass %) PSCL Text FAO texture class (see Appendix 6) BULK Single Bulk density (kg dm-3) TAWC Integer Available water capacity (cm3 cm-3 102 or vol%, -33 kPa to -1.5 MPa) CECS Single Cation exchange capacity (cmolc kg-1) of fine earth fraction BSAT Integer Base saturation as percentage of CECsoil ESP Integer Exchangeable Na as percentage of CECsoil CECc Single CECclay, corrected for contribution of organic matter (cmolc kg-1) PHAQ Single pH measured in water TCEQ Single Total carbonate equivalent (g C kg-1) GYPS Single Gypsum content (g kg-1)




ELCO Single Electrical conductivity (dS m-1) TOTC Single Organic carbon content (g kg-1) TOTN Single Total nitrogen (g kg-1) ECEC Single Effective CEC (cmolc kg-1) ALSA Integer Exchangeable Al as percentage of ECEC

Notes: 1) The soil components that occur within a SOTER unit are numbered sequentially,

starting with the spatially dominant one. The sum of the relative proportions of all component soil units is always 100 per cent. This total will also include a number of unnamed ‘impurities’, commonly in excess of 15 to 30 percent of the map unit (Landon 1991 p. 16-17; Marsman and de Gruijter 1986).

2) Each map unit in the geographic database has a unique identifier (NEWSUID) consisting of the country ISO code (ISOC) and the SOTER unit-ID (SUID); this primary key provides a link to the attribute data for the constituent terrain, terrain component(s) (TCID) and soil components (SCID) (see Figure 1).

3) Tables with the same structure have been prepared for the DOMINANT soil unit only, by depth layer (i.e., for layer D1, see for example table SOTERsummaryFile_T1S1D1) to facilitate visualization using GIS, as example only. Comprehensive studies, however, should consider the full map unit composition and depth range to 1 m.

4) A limited number of records may contain a negative value (-9); this indicates that it has not yet been possible to plug the corresponding gaps using the current taxotransfer scheme due to a lack of measured data in WISE. Whenever possible, virtual profiles in SOTER should be replaced with real, measured profiles after which new secondary data may be generated. A value of ‘-8’ is used for water bodies or SOTER unit ‘KNsn1’.

5) Property estimates are depth-weighted values, per 20 cm layer up to 1m depth and per 50 cm from 1 to 2 m (derived soil properties for 100 to 200 cm, however, are not included in the present secondary database, see text).



Appendix 5. Structure of table SOTERsummaryFile_Prop The field definitions in this table are identical to those used in SOTERsummaryFile. The main difference is that derived data for each soil component of a given SOTER unit are now listed in a single row (horizontally); data for a given layer are preceded by a flag for this layer. For example, field D1_BULK presents derived values for bulk density for layer D1 (0-20 cm), whereas D2-BULK holds data for layer D2 (20-40 cm) and so on. Using this file format, it is easier to query properties of the individual component soil units of a SOTER units using GIS. However, results can only be shown for one soil component, by SOTER unit, at a time (e.g. for TCID=1 and SCID=1).

Structure of table SOTERsummaryFile_Prop


ISOC Text ISO-3166 country code (1994) SUID Integer The identification code of a SOTER on the map and in the database NEWSUID Text Globally unique map unit code, comprising fields ISOC plus SUID TCID Integer Number of terrain component in given map unit SCID Integer Number of soil unit within the given SOTER unit PROP Integer Relative proportion of SCID in given SOTER unit CLAF Text FAO-Unesco Revised Legend code PRID Text Profile ID (see table SOTERunitComposition) Drain Text FAO soil drainage class D1_TopDep Integer Upper depth of layer D1 (0-20 cm) D1_BotDep Integer Lower dept of layer D1 D1_varx Variable Values (e.g., varx is ORGC, BULK, Clay) for layer D1 D2_TopDep Integer Upper depth of layer D2 (20-40 cm) D2_BotDep Integer Lower dept of layer D2 D2_varx Variable Values (e.g, varx is ORGC, BULK, Clay) for layer D2 … … … D5_TopDep Integer Upper depth of layer D5 (80-100 cm) D5_BotDep Integer Lower depth of layer D5 D5_varx Variable Values (e.g, varx is ORGC, BULK, Clay) for layer D5

Note: A table with the same structure has also been prepared for the DOMINANT soil unit only (i.e., TCID= 1 and SCID=1) to facilitate visualization using GIS, as example only (see table SOTERsummaryFile_PROP_SC1). Comprehensive studies, however, should always consider the full map unit composition and depth range.



Appendix 6. Soil textural classes

Soil textural classes (PSCL) are in accordance with revised SOTER criteria (Figure 5). The following abbreviations are used: C–coarse, M–medium, Z–medium fine, F–fine and V–very fine. Further, the symbol u is used for undifferentiated (i.e., C + M + F + Z + V). In addition, all Histosols data have been flagged as consisting of organic materials (O) even though this may not always be the case for all horizons/layers, in a strict taxonomic sense (see FAO 1988 , p. 39)

Figure 5. SOTER soil texture classes (Source: CEC 1985)



Appendix 7. Installation The derived soil data and GIS-files are presented in one single zip file: SOTWIS_Kenya_UpperTana_v1.zip. By default, this compressed file will be unzipped to folder X:\SOTWIS_Kenya_UpperTana, where X is the actual location (i.e. folder). This new folder will contain: - A Readme1st file and the documentation (ISRIC Report 2010/0X) - The project file (SOTWIS_Kenya_UpperTana_v1.mxd) with metadata

(SOTWIS_Kenya_UpperTana_v1.mxd.xml) - Two subfolders:

- GISfiles with the shape and selected layer files files. - SOTWIS with the derived soil data in MSAccess® format

(SOTWIS_Kenya_UpperTana_v1.mdb). The GIS project file (*.mxd) includes several derived data sets for the top layer (0-20 cm) of the dominant soil unit of each SOTER unit (TCID=1, SCID=1), as examples. Actual data applications should consider the full map unit composition, in terms of component soil units, and depth range; see text for details. The dataset has been created using MS-Access® and ArcGIS9/ArcMap9.3®; the shapefiles may also be accessed using ArcView3.3®.

ISRIC - World Soil Information is an independent foundation with a global mandate, funded by the Netherlands Government, and with a strategic association with Wageningen University and Research Centre. Our aims: - To inform and educate - through the World Soil Museum, public information, discussion and publication - As ICSU World Data Centre for Soils, to serve the scientific community as custodian of global soil information - To undertake applied research on land and water resources